Update: This project recently got backed by SAFE and LonghashX and is part of Consensys Fellowship

Problem and Solution

Transactions in web3 have always been pretty primitive and straightforward, there is no easy way to chain them, schedule them, automate them, or batch them all together.

With Epoch Protocol we are trying to bring in Smart Contract Automation, transaction batching, transaction chaining, reading on-chain and off-chain data, and a lot more to take what's possible with the transactions to the next level.

All this started with a simple need, not having to claim poolTogether rewards after every 60 days (They claw them back after 60 days). From there on we iterated a lot on the type of product we needed to build, what was feasible, and how to solve various problems.

Epoch protocol in its initial version was an open market, where anyone could come and execute a job for you by clicking on a button, then it became a node for calling functions on-chain, and by this time account abstraction-related docs dropped by and it all made sense.

With account abstraction, we could do all the things we set out to do while being pluggable in the ecosystem, although we had to solve a lot of problems like nonce management, data availability, adding additional data without breaking signed data, chaining tasks, and a lot more.
We have solved almost all of these issues and working our way through the few remaining, but it's the most fun, solving never been solved before challenges and building something that expands the capability of this amazing ecosystem.

Technical Implementation

Apart from normal fundraising and talking to investors, I undertook designing technical architecture, solving critical problems, and then consequently implementing them.

I worked around the following things -

Technical architecture and solution to transaction automation using account abstraction.
Modifying Etherspot's Transaction bundler implementation to help us listen to on-chain events, multiple mempool to keep transactions for the future and transactions to be sent now separately and lastly sending transactions that a ready to be sent.
Wrote multiple smart contracts
1. Modified Smart contract wallets to skip for the nonce check-in signature, but still prevent replay attacks.
2. Registry contract to keep track of all the tasks and their on-chain conditions.
3. Custom conditional contracts.
4. A byte manipulation library to get data from abi-encoded bytes without knowing its type as well as overwriting it.

Critical problems solved

Problem of nonce
1. Since epoch is all about transaction automation, the sequence of transactions can not be determined and the user's signature can not be changed.
2. The nonce of the transaction has to come dynamically without breaking the signature.
3. Nonce can not be skipped entirely either because it will lead to a replay attack.
The problem of data availability
1. Certain transactions that are supposed to be executed in the future will rely on data from the future.
2. That data needs to be part of transaction without breaking signature.

Web3 Automation with Account Abstraction